Bulimia Nervosa (BN) is a serious disorder characterized by overvalued ideas concerning body weight and shape and disinhibited eating patterns consisting of extreme dieting followed by binge eating and purging. BN occurs in approximately 1-2% of the adolescent population, while clinically significant bulimic behaviors or partial BN (specifically defined as binge eating and purging on average at least once a week for six months) occur in an additional 2-3%. However, other than frequency of binge eating and purging behaviors there appear to be few differences in the clinical severity of those with full syndrome BN and those with partial BN. Although bulimic behaviors typically onset during adolescence, there are only two published randomized clinical trials (RCTs) examining treatments for this age group. Results from our own recently published RCT of adolescent BN using a form of family-based therapy (FBT-BN) has shown that FBT-BN is superior to individual therapy for adolescents with BN. Also, recently published case series reports suggest that CBT adjusted for adolescents (CBT-A) is a promising treatment for bulimia in this age group. The Specific Aims of the proposal are: Aim 1: To compare the efficacy of CBT-A to FBT-BN in decreasing binge eating and purging episodes in adolescents with BN and partial BN (defined as binge eating and purging at least one time per week for the past six months). We predict that subjects randomized to CBT-A will achieve abstinence (as assessed by the Eating Disorder Examination) at significantly higher rates than those who receive FBT- BN. Aim 2: To explore potential moderators of outcome. We predict that symptom severity at baseline will moderate outcome; specifically, those subjects with greater symptom severity (higher EDE subscale scores) will have better outcomes with CBT-A than FBT-BN. Aim 3: To explore potential mediators of outcome. We predict that early changes in the restraint subscale of the EDE-Q) will mediate outcome in both treatments. Aim 4: If there are no differences between the two specific treatments, we will explore abstinence differences between specific treatments (CBT-A and FBT-BN) compared to a non-specific treatment (Supportive Psychotherapy - SPT) in order to examine nonspecific treatment effects (passage of time and nonspecific therapeutic influences). To accomplish these aims the following two-site study is proposed: 158 adolescents (79 at each site) between the ages of 12-18 years who meet DSM- IV criteria for BN or partial BN will be randomized to CBT-A (66 subjects), FBT-BN (66 subjects) or SPT (26 subjects), a non-specific treatment for BN for a 20 session/six month treatment period. The study is powered for the analysis of the primary aim of comparing the two specific treatments (FBT-BN and CBT-A); however, we are randomizing a smaller cohort to SPT to allow a post-hoc examination of non-specific treatment factors should no differences be found between the two specific treatments.